EnglishframeworkID: 1bc193a3
OthersideAI/self-operating-computer
Framework that lets multimodal models view the screen and choose mouse/keyboard actions to operate a computer. Supports GPT-4o/4.1, o1, Gemini, Claude, Qwen-VL, and LLaVA/Ollama modes; installable via pip as `self-operating-computer`.
Category
Automation / agents
Provider
GitHub
Your Library
Editorial Review & Decision Guide
Best For:
- Target Audience: Developers
- Topic focus: Automation (specializing in agents, automation, developer-tooling, workflow)
Access Recommendation: This project is currently flagged for "Write Content" in our workflow. Check our AI review details below before opening the repository.
AI Workflow Audit Data
Key Takeaway: A foundational open-source example of computer-use agents: simple human-like I/O loop, useful for explaining screen-reading + action-taking automation and its current reliability limits.
Priority: P2Suggested Action: Write Content
agentsautomationdeveloper-toolingworkflow
Open Resource ↗